> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://poly-ai.ferndocs.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://poly-ai.ferndocs.com/_mcp/server.

# Tutorial: Edit voice settings

> PolyAcademy Level 1 – Select a voice, configure tuning, and set up your disclaimer.

**Lesson 4 of 6** – Your agent has a personality and knows things – now give it a voice. You'll pick how it sounds and adjust clarity and consistency.

> **Lesson**
>
> Level 1 · Beginner · 10 min

Complete this after behavior is stable but before call testing. The voice is what turns your chat-only agent into a real voice assistant.

## Where this lives

Navigate to **Voice > Settings**

![set-up-project](/_fern-files/_img/e440c75dd0094f8843c9045d797da1269bcdbc51f2a0a4d2aca30b809dac39c9.webp)

You will see two configurable sections:

#### Agent

The primary speaking voice

#### Disclaimer

Used only for legal or informational disclosures (if enabled)

## What you are configuring

Each voice entry has three layers:

#### Voice selection

Who the agent sounds like

#### Weighting

If multiple voices are used, how they're distributed

#### Voice tuning

Stability, clarity, and speed adjustments

## Configure your Agent Voice

#### Select the Agent Voice

In the **Agent** section:

* Click **Change** to open the [Voice Library](/agent-studio/voice/voice-design/voice-library)
* Browse voices using the **Explore** or **Favorites** tabs
* Filter by **Language**, **Region**, and **Gender**
* Preview each voice using the play button or enter custom text

Choose a voice that:

* Matches your audience's expectations
* Sounds neutral and professional
* Remains pleasant over repeated listening

Once selected, click **Select**.

#### (Optional) Add multiple agent voices

You can add more than one agent voice and split usage between them.

* Click **New voice**
* Select an additional voice
* Assign a percentage weight (for example, 70% / 30%)

#### When to use multiple voices

* You want subtle variation across calls
* You are testing two voices in parallel
* Different voices for different scenarios

#### Configure voice settings

Next to each voice, click the **Settings (gear)** icon to open advanced tuning.

#### Stability (%)

Controls how consistent the voice sounds over time. Higher values keep tone and delivery more uniform across responses.

* Higher = more predictable, less expressive

**Recommended:** 0.7 or above for reliable performance.

#### Clarity and similarity (%)

Controls how close the generated voice is to the original recording. Higher values sound more realistic but may be less resilient to unexpected content.

**Recommended ranges:**

* Dynamic conversations: 80–95
* Scripted or repetitive content: 95–100

Click **Done** to save settings.

#### Configure disclaimer voice (if enabled)

If your agent uses a disclaimer:

* Expand the **Disclaimer** section
* Select a voice (can be the same or different from the agent)

**Best practices:**

* Use a clear, neutral voice
* Avoid expressive or casual tones
* Keep speed slightly slower than the main agent

**Example disclaimer text:**
"This call may be recorded for quality and training purposes."

#### Publish changes

Once satisfied:

* Click **Publish**
* Changes will apply to the selected environment

#### Check your understanding

#### Your agent plays a recorded legal disclosure before every call. Which section of the Agent Voice page handles this?

* Agent – it controls all audio played during the call
* Disclaimer – it's designed for legal or informational pre-call recordings
* Behavior – legal requirements should be enforced there
* Environment settings – disclaimers apply per environment

**Answer:** Disclaimer – it's designed for legal or informational pre-call recordings

The Disclaimer section is designed for legal or informational recordings separate from the agent's main voice. Configure it only if your agent actually uses a disclaimer – otherwise leave it unconfigured.

## Verify your voice settings

After publishing, make at least one test call and listen to:

#### Greeting

* Voice sounds clear and welcoming
* Pacing feels natural
* No awkward pauses

#### Mid-call responses

* Voice remains consistent across turns
* Responses flow naturally
* No sudden changes in tone or speed

#### Longer responses

* Voice maintains professionalism throughout
* Instructions are clear and easy to follow
* Pacing allows for comprehension

> **Tip**
>
> **Success criteria:**
>
> * Voice sounds consistent throughout the call
> * No awkward pacing or unnatural pauses
> * Audio matches expectations across all conversation turns
>
> If the voice quality is not as expected, adjust **stability first**, then speed.

#### Check your understanding

#### If the voice sounds off after publishing, what should you adjust first?

* Speed, then pitch
* Stability first, then speed
* Switch to a different voice model
* Pitch first, then stability

**Answer:** Stability first, then speed

Stability controls how consistent and predictable the voice sounds – it's the most impactful dial for basic quality issues. Adjust it first, then fine-tune speed if needed.

## Try it yourself

#### Challenge: Choose settings for a hotel concierge

You are configuring voice for a luxury hotel concierge agent. Decide and explain:

1. Should stability be higher or lower – and why?
2. What clarity range would you use for natural guest conversations?
3. Would you use a single voice or multiple voices for this agent?

#### Hint

For a luxury hotel, consistency and tone matter a lot. Think about what "professional" sounds like in audio – and what settings support that.

#### Example solution

1. **Stability:** High (0.8 or above). Luxury guests expect a consistent, predictable tone – not expressive variation.
2. **Clarity:** 80–90. This range keeps the voice natural and conversational without being brittle to unexpected phrasing.
3. **Single voice.** Multiple voices introduce inconsistency, which conflicts with a luxury brand expectation. Add a second voice only if you're specifically A/B testing.

#### Check your understanding

#### When should you configure a disclaimer voice?

* Always – it's a best practice for every agent
* Only if your agent uses a disclaimer (e.g. "This call may be recorded")
* Only on enterprise plans
* For each environment separately

**Answer:** Only if your agent uses a disclaimer (e.g. "This call may be recorded")

Only configure a disclaimer voice if your agent actually uses one. If your agent doesn't include a disclaimer, this section can be ignored entirely.

## Go deeper

The Voice section of the Help Center covers everything from choosing voices to advanced multi-voice setups:

#### [Voice overview](/agent-studio/voice/overview)

Full reference for all voice configuration options

#### [Voice Library](/agent-studio/voice/voice-design/voice-library)

Browse and preview all available voices

#### [Choosing a good voice](/agent-studio/voice/voice-design/choosing-a-good-voice)

Tips for picking the right voice for your use case

---

#### [← Previous: Add a simple topic](/academy/poly-academy/poly-academy-level-1/add-a-simple-topic)

Lesson 3 of 6

#### [Next: Environments & versions →](/academy/poly-academy/poly-academy-level-1/environments)

Lesson 5 – learn to test safely before going live